A Negotiator’s Challenge to AI: Let’s Talk

A colleague, fellow consultant, and friend - real-life MIT rocket scientist and process expert Stephen Peele Sr. recently sent me an article and a simple question:

“Tom, is this true?”

The article, from The AI Journal, summarizes seven takeaways from the 2025 AI Negotiation Summit, as interpreted by MIT Sloan professor Jared Curhan.

Here are a few standout findings:

• AI bots described as “warm” reached more deals, created more value, and scored higher on subjective outcomes than bots described as “dominant.”

• In transcript analysis, AI outperformed humans in interpretation accuracy. Judges agreed with AI 68% of the time.

• AI could help people who avoid negotiation, like those buying a car or negotiating salary, by lowering anxiety and offering training.

It’s a compelling read. And while I’m not an expert in AI, I do know negotiations. So, I’ll do my best to answer Stephen’s questions.

Here’s my perspective, along with a challenge.

1. Warmth works, but it’s not the whole picture

When Professor Curhan says “warmth beats dominance,” that strikes a cord. The idea is at the heart of the book I co-authored, Negotiating High-Performance, Focused Partnerships. It emphasizes that the best outcomes come from collaboration, cooperation, and joint problem-solving. Game theory backs this up. So does my experience.

In an upcoming Forbes article, I compare successful negotiators to musicians. Both need a trained ear and an awareness of tone, cadence, tempo. Set the wrong tone, and everything else falls apart. If AI can establish a warm tone, that’s promising.

But warmth alone isn’t enough.

Skilled negotiators shift between fostering cooperation and applying pressure. We read the room, listen between the lines, adjust when the tempo changes. Sometimes, the moment calls for a well-timed pause. Other times, it calls for a nudge a firm stand.

Can AI do that yet? Hard to say, but I’ve got my doubts.

2. Outcome vs. Interpretation

AI might be able to analyze words on a page. It can count how often certain phrases show up or maybe even flag indicators of friction or alignment. But negotiation isn’t just a language puzzle. It’s a results game.

It’s about outcomes: stronger partnerships, competitive advantage, flexibility, long-term success.

I don’t walk out of a negotiation patting myself on the back because we “sounded aligned.” I look at whether we moved the needle. Did we solve a real problem? Did we unlock flexibility, protect long-term interests, or strengthen the relationship so the next conversation goes even better?

The answers to those questions don’t always show up in a transcript.

It’s the moments between the words at matter. It’s the tone someone uses or a slight hesitation that tells a trained negotiator where things really stand. Those moments shape what happens next.

If AI is judging success by whether people appear to agree, or whether language aligns with intent, that’s useful. But it’s not the full picture. The full picture is whether people walk away with a deal that holds.

3. AI can lower the barrier to entry

Here’s where I think AI has real potential. Not everyone likes negotiating. Some people dread conversations like salary talks or car purchases.

I’ve coached people who are sharp, strategic thinkers, but freeze up when it’s time to ask for what they need. They worry about sounding pushy, or stall when things go off script.

Maybe AI can help folks build muscle memory, hear themselves in action, and spot patterns in how they respond. Tools like training simulations or role-play assistants could make negotiation feel less intimidating and more approachable.

But let’s be clear: that kind of training shouldn’t be confused with the work negotiators do.

The kinds of negotiations I’m most familiar with – organizational change, labor contracts, complex partnerships – don’t follow scripts. They’re full of personalities, relationships, risk, and long-term consequences. They evolve minute by minute. They require you to listen and adapt.

AI may help people prepare for those moments. But in the room? When the deal could go sideways? That’s still very much human terrain.

The Posey AI Negotiation Challenge

So, here’s my ask:

I’d love to test this in the field. I’m looking for a partner willing to give me access to one (or several) of these AI negotiation bots. Together we can set up three complex, real-world scenarios. We can use the same protocols from the recent competitions or tweak them based on what matters in real negotiations.

Then let’s see what we learn.

I’m not here to beat the machine. I’m here to understand what it can, and can’t, yet do.

Anyone game?